More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Systematic Debugging: Diagnose Before You Fix

    1

    Debug by diagnosis, not by guessing: establish the root cause with evidence, localize across boundaries, test one hypothesis at a time, and prove the fix stays fixed.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Systematic Debugging: Diagnose Before You Fix

    Systematic Debugging: Diagnose Before You Fix

    Example session with this skill installed

    A deploy pipeline fails intermittently — about one run in twenty — and nobody knows which stage is guilty. Diagnose it without guessing.

    • Read your context and instructions
    • Compiled the systematic debugging

    Read the evidence first: the failing runs all showed the same error at the signing stage, but the log lines around it pointed upstream — a variable that was present in step two and absent in step three. Reproduction was measured, not shrugged at: a loop ran the pipeline twenty times and failed once, matching the reported one-in-twenty rate, with the exact inputs recorded. Recent changes were checked next: the build script had been edited three days earlier, and the diff showed the variable being set in a scope that step three could not see. Localization was one evidence run, not an hour of reading code: at each component boundary, what went in and what came out was logged, and the run showed exactly where the good data became bad — the build script dropping the variable between step two and step three. A working-versus-broken comparison confirmed it: the last known-good version set the variable at pipeline scope, and the current version set it at step scope. The hypothesis was written in one sentence and tested with the smallest possible change — move the assignment to pipeline scope — and the loop ran twenty times with zero failures. The failing test was written first as a regression test: the pipeline now asserts the variable is present at every stage boundary before signing. The existing suite still passes, and the original symptom is gone in the real environment, not just in the test. The ledger records the diagnosis, the single fix, and the verification — and the one thing that was not verified is named plainly: the signing stage's retry behavior under a genuinely missing credential was not tested, and it is a separate investigation, not a footnote to this fix.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    No fix without a diagnosed root cause. Not a probable cause, not a plausible story — a cause established with evidence. Symptom fixes are how one bug becomes a family of bugs: the retry loop that masks the broken invariant, the null check that hides the uninitialized state, the catch-all that swallows the real error. This skill covers the full discipline: reading the evidence first, reproducing reliably, checking what changed, tracing the data to its source, bisecting the system across component boundaries, comparing working versus broken, and testing one hypothesis at a time with the smallest possible change. It covers fixing at the source with a failing test written first, verifying fully, and the three-strikes rule — when three fix attempts have failed, the problem is probably in the design, and it is time to question the fundamentals with the human instead of attempting a fourth fix.

    It diagnoses; it does not fix your code for you — the fix is yours, with the failing test proving it.
    Intermittent failures need a loop and a failure rate; if a failure cannot be reproduced, the skill says so and gathers more data instead of guessing.
    The three-strikes rule is a design discussion with the human, not a fourth fix attempt.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Requires

    Git

    Permissions

    Terminal / Shell
    Listed3 days ago

    What's inside

    Frequently Asked Questions

    Because pressure makes guessing feel productive and it is not. Thrashing is the slowest debugging strategy there is — a diagnosed cause with one minimal test is faster than three guesses.

    Then you cannot verify a fix. Say so, gather more data — a loop, a failure rate, the exact inputs — instead of guessing.

    After three failed attempts, when each fix reveals a new problem elsewhere and each fix needs increasingly heroic refactoring. That is the signature of a design problem, not a detail problem.

    New in Testing & Debugging